The representationist maintains that an experience represents a state of affairs. To elaborate, a stimulus of one’s sensorium produces, according to her, a "phenomenal composite" made up of "phenomenal properties" that are the typical effects of certain mind-independent features of the world, which are thereby represented.¹ It is such features, via their phenomenal representatives, of which the subject of an experience would become aware were she to engage in introspection. So, one might ask, what state of affairs would be represented by an illusory experience, that is, one to which no state of affairs in the vicinity of its subject corresponds? The answer, according to the standard defense of representationism (SD), is the same state of affairs that would obtain in its subject’s surroundings if it were true.

Representationism is held in opposition to "phenomenism," which treats some of the features of a phenomenal composite, whether veridical or not, as being logically independent of the features of its stimulus, that is, as uniquely mental media.² Here, introspection reveals the features of experiences themselves, not only of mind-independent objects. As Ned Block has put it, a state of affairs is partly depicted using "mental paint."³ On this view, an illusory experience will be marked by a feature of the medium of its representation that will distinguish it from an experience to which a "present" fact, to use Hume’s term, corresponds. Phenomenists appeal to such experiences, e.g., those involving after-images or so called "impossible figures," to demonstrate that an experience must transcend its representational content.

Before addressing this issue, the representationist must answer the question of how the subject of an experience, who is aware of its illusoriness, could be representing a state of affairs without believing that her experience is veridical.⁴ The representationist maintains, however, that to represent a state of affairs it is not necessary to think of it as being in the vicinity: to borrow again from Hume, one can represent an "absent" state of affairs. As long as the subject of an experience is disposed to act as if a state of affairs obtains in her surroundings, it should be taken as being represented.⁵ This response only invites the rebuttal that one can hardly be disposed to act as if a state of affairs is in the vicinity without believing that it is present. It would be better here, I believe, to appeal to Michael Tye’s distinction between a "conceptual and a non-conceptual representation" and insist that while experiencing the latter, the subject of an illusory experience would believe that the state of affairs that would make her experience true is present: as long as one is ‘under the illusion’, what one is experiencing will be different than what one may eventually realize that one is experiencing.⁶ But I shall set this issue aside, as I am more concerned with the just mentioned deeper problem facing an advocate of SD: explaining the "suspiciousness" of false experiences of certain types on the assumption that they are phenomenally indistinguishable from veridical experiences.⁷

It seems that if a subject is aware of the illusoriness of an experience, then she must be distinguishing it from experiences that could represent her surroundings: her incredulity can only be explained by her awareness that her experience transcends what she can represent. To put it another way, how could a subject conceptualize a non-veridical experience as false if its content were wholly representational, that is, if it were lacking a mind-dependent feature an awareness of which would prompt the realization that what is being represented is an absent state of affairs? SD seems inadequate here. To solve this problem, Robert Schroer appeals to the fact that an illusory experience would be made up of an unusual combination of mundane phenomenal properties. This discrepancy is made to account for its suspiciousness: compared to a typical veridical experience, its components are assembled in violation of the laws of "folk" science, which its subject presumably realizes. The neural mechanisms that would produce a representation in conformity with these laws may operate asynchronously, thus allowing for strange combinations of the phenomenological properties of which it is composed.⁸ According to Schroer’s alternative, then:

The components of a non-veridical experience are representations ‘adding up’ to the representation of an "unnatural" state of affairs (SA).

The treatment of the experience of after-images entailed by SA is consistent with SD, since the latter only requires that a false experience be qualitatively identical to a possible experience, not to one that conforms to folk science. SA explains, however, why some experiences are atypical, striking one as odd, for which SD fails to account. SA diverges from SD, however, when it comes to handling experiences whose truth-conditions violate the law of non-contradiction. As Christopher Peacocke has pointed out, SD could not account for these experiences, if indeed they are possible, since they could not be veridical: since there could not be a contradictory state of affairs, such experiences are necessarily unrepresentative of anything.⁹ Thus, we must first determine whether or not they could occur.

Schroer rightly maintains that Peacocke’s cases, which deal with experiences whose typical "characterizations" seem to transcend their representational content, do not involve contradictory truth-conditions.¹⁰ For his part, he cites the waterfall illusion, in which the rocks alongside a waterfall are supposedly represented as moving sans a representation of them changing location, as an experience that lacks a veridical counterpart.¹¹ Since motion just is change of location, one wonders how things could appear this way. It seems more likely that we are dealing here with an ambiguous figure, something akin to the "duck-rabbit," which does not appear simultaneously as a drawing of a duck and a rabbit, but as a drawing of one or the other depending on the viewer’s focus. Tye adopts this strategy to rebut one of Peacocke’s putative counterexamples to SD: the case of an object coming to occupy a larger space in one’s field of vision as the distance between it and oneself is reduced.¹² To account for the phenomenal difference here, Tye posits "viewpoint dependent" properties, such as how an object would appear from here, as aspects of an experience’s representational content. If such properties are contained within a phenomenal composite, then there is a state of affairs to which the experience of a so called "impossible figure" corresponds, viz., one that would be seen in one way from a certain perspective and in another contrary way from a different viewpoint.

I shall proceed, however, as if this case is a counterexample to SD. Can it still be handled without completely abandoning representationism? Schroer believes that the representationist can account for it, provided that she distinguishes between the representativeness of a phenomenal composite and the representativeness of each of the phenomenal properties of which it is composed. Despite having no mind-independent correlate as whole, Schroer argues that the phenomenal properties of the waterfall illusion, taken individually, are still representations, "their status as represented features (being) independent of the veridicality (or the possible veridicality) of the experiences (in which they are contained)."¹³ Thus, our sensorium is taken here to be capable of producing phenomenal composites that not only violates folk science, but the laws of logic as well!

This move, I believe, is not in keeping with the basic tenet of representationism. A sensory experience, on this view, is supposed to represent a possible state of affairs, not just a collection of (mind-independent) properties. Any false experience should, thus, be understood by a representationist as a representation suspected of failing to represent what could but does not in fact obtain (assuming that it is possible to be representing something while failing to believe that it is in the vicinity). Its components are supposed to be representations ‘adding up’ to a fuller representation, as is the case with a veridical experience, not just an incoherent composite. There should not be a discrepancy between what the phenomenal properties of veridical and non-veridical experiences can ‘amount to’. Moreover, if, as Frege held, the terms of a meaningless sentence are themselves meaningless, the phenomenal properties that make up an incoherent composite must be unrepresentative of anything as well.¹⁴ What holds true of our written and spoken systems of communication should also hold true of the "language of thought": the meaninglessness of a mental representation rendering its parts meaningless as well. Just as ‘boredom’ cannot meaningfully appear as the subject of a sentence whose predicate is ‘is laughing’, a thing’s motion cannot appear within a phenomenal composite that also includes a representation of it staying in one place.

In sum, SA when applied to after-image experiences serves as an addendum to SD, explaining strange, but logically coherent, phenomenological composites. As far as so-called contradictory experiences are concerned, the representationist seems better off arguing that they do not really occur, as SA does not succeed in explaining why they should be considered, either in whole or in part, representations.

1. Robert Schroer, "After-Images, Phosphenes, and Other Movies in the Psychedelic Theater," (manuscript, presented at the 2002 APA Central Division Meeting), pp. 7-8. Other contemporary representationists include Fred Dretske, (Naturalizing the Mind, Cambridge, MA: MIT Press, 1995), Michael Tye (Consciousness, Color, and Content, Cambridge, MA: 2000), and Gilbert Harman ("The Intrinsic Quality of Experience," in Philosophical Perspectives 4 ed. J. Tomberlin, Northridge, CA: Ridgeview, 1990)
2. Contemporary phenomenists include Ned Block ("Mental Paint and Mental Latex," in Philosophical Issues 7 ed. E. Villenueva, Northridge, CA: Ridgeview, 1996), Christopher Peacocke (Sense and Content, Oxford: Oxford University Press, 1983), and Paul Boghossian and David Velleman, ("Color as a Secondary Quality," Mind 98 (1989): 81-103).
3. Block, op. cit., pass..
4. Robert Schroer, op. cit., pp. 3-4.
5. Ibid., p. 4. This idea was first advanced by Gareth Evans. See The Varieties of Reference, Oxford: Oxford University Press, 1982, pp. 121-129.
6. Tye, op. cit. pp. 75-6.
7. Schroer, op. cit., pp. 4-6.
8. Ibid., p. 8.
9. Peacocke, op. cit., pp. 12-26. Tye (op. cit., pp. 85-6) maintains that the content of an after-image experience is "necessarily inconsistent", since the after-image "appears in front of a yellow surface (and) it does not appear to be there." But, he goes on to say that ‘appear’ is ambiguous here, first being used in the "nonconceptual" and then in the "conceptual" sense. Thus, an after-image experience turns out to be consistent after all: each one of its phases is internally consistent, though they fail to cohere with each other. The subject does not simultaneously see the after-image as being and not being in a certain location. The analogy that he offers to clarify his position only serves to reinforce this point. Tye likens the subject of an after-image experience to a judge who listens attentively to the testimony of a witness but ultimately rejects its as false based on other evidence that he finds more credible. Despite the conflict between the pieces of evidence before him, the judge’s beliefs are perfectly consistent: he does not believe the witness and disbelieve her. His assessment is that her testimony is false, he has heard it without lending it credence. In the same way, once the subject of an after-image experience applies the "conceptual or epistemic sense of appearing" to his experience, it is transformed. Things no longer appear as they did; he is no longer ‘under an illusion’, as we say.
10. Schroer, p. 9, n. 9.
11. Schroer, p. 9.
12. Tye, op. cit., pp. 78-79. Viewpoint relative properties are visually represented, according to Tye, given the correlation between the sizes and configurations of retinal images and the ways in which things are viewed.
13. Schroer, p. 10.
14. Frege writes: "it is enough if the proposition (Satz) as a whole has a sense; it is through this that the parts also acquire their content," in The Foundations of Arithmetic, trans. J. L. Austin (London: Blackwell, 1953) p. x. For discussions of this view, see Michael Dummett, "Frege’s Philosophy," pp. 94-5 in Truth and Other Enigmas (Cambridge, MA: Harvard University Press, 1978) pp. 87-115 and Jeremy D. B. Walker, A Study of Frege, p. 71 and chs. 8 and 9 (Ithaca, NY: Cornell University Press, 1965).